A field experiment was conducted to evaluate weed control systems in a conservation tillage rotation of grain sorghum [Sorghum bicolor (L.) Moench.] – cotton (Gossypium hirsutum L.) – wheat (Triticum aestivum L.). Herbicide systems included fall and spring/summer inputs of high and low intensity. Tillage regimes were no-till (NT) and reduced-till (RT) systems; the latter included fall primary tillage followed by spring stale seedbed planting. Both tillage systems utilized controlled traffic lanes and wide, raised beds. Effective johnsongrass [Sorghum halepense (L.) Pers. # SORHA] control required intense herbicide inputs at one or both application periods, i.e., in the fall and/or spring/summer. Grain sorghum and cotton yields for the most intense weed control system, which included high inputs in both the fall and spring/summer, were not superior to systems that included high inputs in only one of the two application periods. Seedling johnsongrass emergence occurred before spring planting in RT (but not in NT) in 2 of 3 yr, and control measures were ineffective. After 3 yr, the predominant weeds were johnsongrass and browntop panicum (Panicum fasciculatum Sw. # PANFA).